Brain Implants Let Two ALS Patients Speak and Sing, Team Reports

Researchers at the University of California, Davis, have used experimental brain implants to translate the brain signals of a man with amyotrophic lateral sclerosis directly into intelligible speech and to let a woman with the disease sing a musical scale, the team reported.

The work was scheduled to be presented at the Society for Neuroscience conference in Washington. The findings have not yet been published in a peer-reviewed journal.

According to reporting on the research, the man is 46 and lost his voice to ALS. The woman is 58, a trained singer whose voice was gradually silenced by the disease.

A step beyond text

Earlier implant systems paired with artificial intelligence restored communication for people who had lost control of their speech muscles through ALS or stroke, but those systems converted brain signals to text before producing speech.

Under that approach, electrodes implanted in the brain detected signals as a person attempted to speak, and software decoded them into phonemes, the sound units that make up words. Words then appeared one by one on a screen, and the participant triggered a cursor click to finish the sentence before a computer-generated version of the person's voice spoke it.

The process produced a slight delay that made it hard for participants to join conversations, since people often rely on the rhythm, timing and emphasis of speech.

The Davis team's newer approach, building on work the group published in 2025 in the journal Nature on real-time expressive speech generation from brain to voice, aims to close that gap.

Independent assessment

Cindy Chestek, a biomedical engineer at the University of Michigan who had no role in the research, told The New York Times that the system is, for the first time, "good enough" for real communication and "good enough to sing."

Researchers have framed the broader effort as a way to help people with ALS, Parkinson's disease and brain injuries communicate again. The field has advanced quickly, with several groups reporting gains in decoding attempted speech from brain activity.

Because the latest results were presented at a scientific meeting rather than published, they have not undergone the peer review that would typically precede formal clinical claims. Independent replication and larger trials would be needed to establish how well the approach works across patients.